Progress in Code Question Answering (CodeQA) depends on benchmarks that reflect what developers actually ask. Yet most existing CodeQA datasets are built from docstring templates or GitHub issues, and rarely capture the diversity of real, repository-level developer queries. We address this gap by releasing a taxonomy of real-world code-assistance queries, grounded in 3,052 authentic user interactions from the RepoChat dataset. We construct the taxonomy via a semi-automated pipeline that combines clustering, LLM-based labeling, and an iterative refinement loop driven by inter-model disagreement, and we compare the result against three prior taxonomy-generation methods (TaxoGen, TaxoCom, TaxoAdapt). Our taxonomy achieves a favorable tradeoff between granularity and usability—86 nodes with 93.2% node relevance, 99.4% path validity, and 81.25% sibling coherence—while remaining concise enough for downstream use. We then apply the taxonomy in two ways: (i) auditing existing Code related benchmarks to surface systematically underrepresented query types, and (ii) generating a synthetic CodeQA dataset whose categories are guided by the taxonomy.

A domain-specific extension of C language known as extended Berkeley Packet Filter (eBPF) has gained widespread acceptance for various tasks, including observability, security, and network acceleration in the cloud community. Due to its recency and complexity, there is an overwhelming need for natural language summaries of existing eBPF codes (particularly open-source code) for practitioners and developers, which will go a long way in easing the understanding and development of new code. However, being a niche Domain-Specific Language (DSL), there is a scarcity of available training data. In this paper, we investigate the effectiveness of LLMs for summarizing low-resource DSLs, in the context of eBPF codes. Specifically, we propose a clustering-based technique to retrieve in-context examples that are semantically closer to the test example and propose a very simple yet powerful prompt design that yields superior-quality code summary generation. Experimental results show that our proposed retrieval approach for prompt generation improves the eBPF code summarization accuracy up to 12.9 BLEU points over other prompting techniques. The codes are available at https://github.com/babangain/ebpf_summ.